# Explaining adoption of AI tools in education: a dual-path model of ethical concern and functional value

## TL;DR

This study explores how people decide to use AI tools in education by considering both their usefulness and ethical concerns.

## Contribution

The paper introduces a dual-path model combining TAM and PMT to explain AI tool adoption in education.

## Key findings

- Perceived usefulness and ease of use positively predict behavioral and continuance intentions.
- Ethical concerns negatively affect behavioral intention, especially for those with higher moral sensitivity.
- Improving usability and ethical transparency can promote responsible AI tool adoption in education.

## Abstract

As Artificial Intelligence-Generated Content (AIGC) tools (e.g., ChatGPT for writing assistance, Midjourney for image generation) diffuse into educational settings, their adoption reflects a psychological interplay between functional appraisals and ethical concerns. This study proposes and tests a dual-path model integrating the Technology Acceptance Model (TAM) with Protection Motivation Theory (PMT), incorporating Perceived Ethical Concern (PEC) and Moral Sensitivity (MS).

Ten latent constructs were modeled: Perceived Severity (PS), Perceived Vulnerability (PV), Self-Efficacy (SE), Response Efficacy (RE), Perceived Ease of Use (PEOU), Perceived Usefulness (PU), PEC, MS, Behavioral Intention (BI), and Continuance Intention (CI). Using structural equation modeling based on data from 589 respondents with prior AIGC experience, we evaluated 14 hypotheses.

The results support the TAM pathway: PU and PEOU positively predict BI and CI. Meanwhile, PMT components operate indirectly; RE and SE influence appraisals by elevating PU and mitigating PEC, whereas PS and PV elevate PEC. PEC shows a significant negative effect on BI and an indirect negative impact on CI. Notably, this negative PEC-BI association is more pronounced among individuals with higher MS.

The findings extend psychological accounts of AI tool adoption by jointly modeling moral appraisal and functional value in educational contexts. Furthermore, the study offers actionable implications for platform design and policy, suggesting that improving usability and efficacy cues while increasing ethical transparency can foster responsible, sustained use.